Model sequence report

The data for the models was derived for the initial dataset dsL:

numID<- 9022 # highest id value (max = 9022)
### Define the data that will populate the model
ds<- dsL %>%  # chose conditions to apply in creating dataset for modeling
  dplyr::filter(id < numID) %.% # 1:9022
  dplyr::filter(year %in% c(2000:2011)) %.% # 1997:2011
  dplyr::filter(sample %in% c(1)) %.% # 0-Oversample; 1-Cross-Sectional
  dplyr::filter(race %in% c(4)) %.% # 1-Black; 2-Hispanis; 3-Mixed; 4-White
  dplyr::filter(byear %in% c(1980:1984)) %.% # birth year 1980:1984
  dplyr::filter(ave(!is.na(attend), id, FUN = all)) %.% # only complete trajectories
  dplyr::mutate( # compute new variables
    age= year-byear, # definition of age to be used in the model    
    timec=year-2000, # metric of time is rounds of NSLY97 in years, centered at 2000
    timec2= timec^2, 
    timec3= timec^3,
    #         timec= age-16, # metric of time is bilogical age in years, centered at 16
    #         timec2= timec^2,
    #         timec3= timec^3,# 
    cohort=byear-1980) %.% # age difference, years younger (unit - 1 cohort away)
  dplyr::select( # assemble the dataset for modeling
    id, sample, race, byear,cohort, # Time Invariant variables
    year,
    age, timec,timec2,timec3, attend)  # Time Variant variables

m0F – Fixed only

m1F

m2F

m3F

m4F

m5F

m6F

m7F

mFa –

mFb

mFc

mFf

mFd

mFe

m1R1 – 1 Random

m2R1

m3R1

m4R1

m5R1

m6R1

m7R1

mR1a –

mR1b

mR1c

mR1f

mR1d

mR1e

m1R2 – 2 Random

m2R2

m3R2

m4R2

m5R2

m6R2

m7R2

mR2b –

mR2c

mR2f

mR2d

mR2e

m2R3 – 3 Random

m3R3

m4R3

m5R3

m6R3

m7R3

mR3f –

mR3d

mR3e

m3R4 – 4 Random

m4R4

m5R4

m6R4

m7R4

Model specification

General case of LCM and model sequence are defined as: